claude-setup

name	claude-setup
description	Diagnose whether Claude Code is installed and authenticated for this local Codex plugin without making a billable Claude call.

Claude Setup

Use this skill when the user wants to check whether Claude Code is available for the Codex Claude companion plugin, or when Claude delegation fails before a task starts.

Run the setup diagnostic with the companion script from the installed plugin root. If the current working directory is not this plugin checkout, do not first try node scripts/claude-companion.mjs; resolve the script relative to this skill file, for example ../../scripts/claude-companion.mjs from the skill directory.

From the plugin root:

node scripts/claude-companion.mjs setup

For machine-readable output, run:

node scripts/claude-companion.mjs setup --json

To print the optional unmanaged-local Codex approval profile and narrow command prefixes without probing Claude or writing files, run:

node scripts/claude-companion.mjs setup-profile

The setup diagnostic is intentionally non-billable. It only checks:

• node --version
• npm --version
• claude --version
• claude auth status --text

If Claude Code is missing, guide the user to install it with:

npm install -g @anthropic-ai/claude-code
claude install stable

If Claude Code is unauthenticated, guide the user to use one of:

• claude auth login --claudeai for Claude subscription accounts
• claude setup-token for long-lived subscription auth in strict bare mode
• ANTHROPIC_API_KEY
• CLAUDE_CODE_USE_BEDROCK=1 with AWS credentials
• CLAUDE_CODE_USE_VERTEX=1 with GCP credentials
• Claude Code apiKeyHelper

The setup command treats ANTHROPIC_API_KEY, CLAUDE_CODE_USE_BEDROCK=1, or CLAUDE_CODE_USE_VERTEX=1 as configured environment auth and reports claudeAuth.status: "env_configured". Claude Code still validates API keys or provider credentials at runtime.

If claude auth status --text succeeds in the user's normal terminal but the plugin setup check reports unauthenticated, treat it as a sandbox/keychain visibility problem instead of a missing login. Ask the user to approve the Claude-invoking command outside the sandbox, or guide them to bare-compatible auth such as claude setup-token, ANTHROPIC_API_KEY, provider credentials, or apiKeyHelper.

For a smoother unmanaged local Codex setup, recommend a one-time profile that routes approvals to the user:

cat > ~/.codex/claude-companion.config.toml <<'EOF'
approval_policy = "on-request"
approvals_reviewer = "user"
sandbox_mode = "workspace-write"
EOF
codex --profile claude-companion

If organization policy forces automatic approval review and denies external Claude disclosure, the plugin cannot bypass that policy. Tell the user that an admin or workspace policy must allow the narrow Claude delegation command prefixes before live Claude calls can run.

If setup is ready but the first live rescue, plan, ui, or review command is denied by Codex or the host approval reviewer, diagnose that as a host permission block, not Claude auth failure. Report the exact narrow command prefix Codex asked to run and stop. The setup-profile output prints prefixes using the current script path, so a repo checkout and an installed plugin cache can show different paths. Do not fall back to a local Codex implementation unless the user explicitly asks to proceed without Claude.

Do not run a prompt through Claude as part of setup.